​The "HH-DQN" project presents a hybrid approach to solving the Pickup and Delivery Problem (PDP) by integrating Deep Q-Learning with hyper-heuristics. The PDP is a complex combinatorial optimization challenge that focuses on planning efficient routes for vehicles to pick up and deliver goods while minimizing factors like cost, time, and distance, all within specific constraints. Traditional heuristic methods often struggle with adaptability across different problem scenarios. This project addresses these limitations by employing a Hyper-Heuristic based Deep Q-Learning framework, which dynamically selects from a set of low-level heuristics during route construction, aiming to enhance solution quality and adaptability.